Motor learning is a fundamental form of learning important for the well-being of many animal species including humans. The importance of learned motor programs is underscored when they are compromised in motor disorders such as multiple sclerosis and ALS. Neural circuit mechanisms of motor learning have been extensively studied; however, precise plasticity mechanisms of distinct neuron types underlying motor learning are not well understood. We address this issue in the motor cortex, a critical brain region responsible for motor learning. The central hypothesis in this proposal is that subtype-specific changes of inhibitory neurons regulate the plasticity of excitatory circuits necessary for motor learning. To directly visualize these plasticity events within the motor cortex during motor learning, we will apply in vivo two-photon imaging chronically in awake mice performing a motor learning task over weeks, focusing on three major neuron types in the motor cortex (principal excitatory neurons, parvalbumin-expressing inhibitory neurons (PV-INs), and somatostatin-expressing inhibitory neurons (SOM-INs)). We recently developed a lever-press task as a motor learning paradigm for head-fixed mice. We found that learning of this task over two weeks induces a novel and reproducible activity pattern in motor cortex excitatory neuron ensembles. This activity change coincided with a turnover of dendritic spines, the major postsynaptic sites of excitatory synapses, on the excitatory neurons (Peters et al. Nature 2014). Following up on these initial findings, this proposal aims to reveal the role of inhibitory circuits in regulating he plasticity of excitatory circuits. In Aims 1&2, we will characterize the activity and synapse number of PV- and SOM-INs during learning. We hypothesize that motor learning transiently increases PV inhibition and decreases SOM inhibition. We will test this hypothesis by chronically imaging the activity of PV- and SOM-INs using GCaMP6f and their axonal presynaptic terminals. In Aim 3, we will test the hypothesis that the decrease in SOM inhibition is important for excitatory synaptic plasticity and learning. We will test this by manipulating SOM-IN activity using optogenetics and examine the effect on learning and dendritic spine turnover. Finally in Aim 4, we will develop additional motor learning paradigms for head-fixed mice, which will be combined with above experiments in the future to test how generalizable our findings on plasticity mechanisms are to various tasks. These experiments combine cutting-edge technologies including chronic high-resolution two-photon imaging, behavioral tasks by head-fixed mice, mouse genetics to label specific neuron types and optogenetics. These experiments will reveal fine-scale circuit plasticity underlying motor learning and also establish paradigm that can be applied to other forms of learning and behaviors in the future.